This is a *FREE Architect Solution* and was developed to classify vegetation risk areas around powerlines combining aerial optical images (Naip), vectors (powerline shapefile) and a point cloud (elevation, also from Naip). 


If you haven't heard about eCognition Architect or Architect Solutions please have a look at these videos before you continue here and have a look at this Architect Solution.

Input data

  • RGBNIR raster image
  • Powerline vector
  • Point Cloud


  • Point Cloud Critical (classified risk areas in PC)

Quick guide:

  1. Download the Example Architect Solution file and unzip it
  2. Open eCognition Developer or eCognition Architect
  3. Open the provided project file (Drag&Drop) from the "data" directory "Power_Line_Project.dpr"
  4. Go to tab --> Architect --> Open Action Library and select the "ActionLibrary" folder to open
  5. This will open the Solution
  6. Now you can start working with the solution

Workflow explanation:

The idea was to develop a guided User Interface to allow the user to configure several variables themselves based on their needs. Those are:

  1. Buffer distance to the powerline to determine the ROI (Region Of Interest)
  2. Max allowed Height to determine which height vegetation needs to have to be considered as "critical"
  3. Min allowed Distance to determine the distance to the powerline where the algorithm should look for "critical" areas

Step 1 --> The workflow starts with the "Create Buffer" section asking the user to define the Power Line vector and the desired buffer distance. After execution the view will update and display the resulting buffered vector our ROI.

Step 2 --> This is followed by the "Classify Trees and Buildings" section, which asks for definition of several layers that are needed in the Rule Set to derive elevation layers (DSM, DTM and nDSM) and an index layer (NDVI) used for classification as well as to create image objects. After execution, you will see, that image objects are created, only within the ROI (depending on previously defined buffer distance). These objects have been classified, most important are the classes that are displayed after execution which represent elevated vegetation (different height classes).

Step 3 --> This is followed by the identification of the critical areas in the "Identify Critical Trees" section where the user is ask to configure two variables (that are explained above, Max allowed Height and Min allowed Distance) to create a user tailored analysis. After execution critical areas are now highlighted in purple.

Step 4 --> Finally the "Export classified PC" section. The user is asked to define the output folder. After hitting the "Execute" button it will create a "results" folder in the defined directory and export the Point Cloud Critical ({:Project.Name}.v{:Project.Ver}.las) in this directory. The points in the PC that are classified as "critical" are not all points within the detected areas, solely those that are higher than the Max allowed Height threshold.


Please go ahead and use this Solution and Rule Set as a blue print or simply go ahead and use it for your projects. Edit it, improve it (it is by far not perfect!) and try to create superior results! This will also work with the FREE eCOGNITION TRIAL VERSION which means, no excuses ;).


Your Answer

Articles in this section

Need more help?

Enter Community Contact Support